In real industrial engineering, job shop scheduling problem (JSSP) is considered to be one of the most difficult and tricky non-deterministic polynomial-time (NP)-hard problems. This study proposes a new hybrid heuristic algorithm for solving JSSP inspired by the tissue-like membrane system. The framework of the proposed algorithm incorporates improved genetic algorithms (GA), modified rumor particle swarm optimization (PSO), and fine-grained local search methods (LSM). To effectively alleviate the premature convergence of GA, the improved GA uses adaptive crossover and mutation probabilities. Taking into account the improvement of the diversity of the population, the rumor PSO is discretized to interactively optimize the population. In addition, a local search operator incorporating critical path recognition is designed to enhance the local search ability of the population. Experiment with 24 benchmark instances show that the proposed algorithm outperforms other latest comparative algorithms, and hybrid optimization strategies that complement each other in performance can better break through the original limitations of the single meta-heuristic method.
An improved deep reinforcement learning approach for the dynamic job shop scheduling problem with random job arrivals
